A test campaign characterizing the flow separation in overexpanded subscale nozzles has been performed in the R2Ch blowdown wind tunnel of the Onera Chalais-Meudon center. As the side load phenomenon is caused by flow separation, side load theoretical models could be improved if we had a better understanding of the aerodynamics of overexpanded nozzles. This need has motivated the present experimental study. Two axisymmetric nozzle models have been tested: one having a truncated ideal contour (TIC), the other being designed with a thrust-optimized contour (TOC) and having the ability to simulate wall film injection. The flow separation description inside these overexpanded nozzles relies on a variety of measurements. Schlieren photographs of the exhaust jets have shown the overexpansion shock structures, and particularly the cap-shock pattern visualized in the jet of the TOC nozzle, which results from the interference between the internal shock and the overexpansion shock. Probings with a Pitot tube in the jet have given pressure values on the centerline, upstream and downstream of the Mach disc. The analysis of mean wall pressure gives an estimation of the incipient flow separation position versus nozzle pressure ratio. These tests have revealed the influence of the film injection on flow separation location. The analysis of unsteady wall pressure measurements shows that the pressure rms levels reach a maximum of 40% of the mean local unperturbated pressure in the shock wave / boundary layer interaction regions of the TIC nozzle, which is higher than pressure fluctuation levels measured in compression ramps. The pressure rms values are about 5 to 8% of the mean local unperturbated pressure in the separated flow zones. Nomenclature C shock L nozzle divergent length Lsep interaction length Maxis Mach number on centerline p pressure pa ambient pressure (in plenum chamber) Paxis pressure on centerline Po pressure at the origin of the interaction pst stagnation pressure p'st stagnation pressure measured by Pitot tube pressure rms X abscissa T triple point I, slip line Introduction For rocket engine applications, the first goal of the propulsive nozzle is to optimize the thrust-to-weight ratio. As the nozzle weight becomes an important aspect, the optimization consists of largely reducing the divergent length while designing a profile evolution with a large expansion ratio. * Research scientist f Technical engineer * Research engineer § Researc scientist Copyright © 2001 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved. This optimization process necessitates a rapid evolution of the profile curvature, especially in the region just downstream of the throat. This induces high speed and high momentum in the core flow, while the flow recompression near the divergent wall ensures high wall pressure levels which is favorable regarding the separation phenomenon. Such a thrustoptimization leads to a high non-uniformity of the flow in the exit plane. As the large truncature of the divergent is the most outstanding characteristic of these nozzles compared to the ideal nozzles, these moderately thrust-optimized nozzles, as the Vikingtype nozzles [1], are called here "truncated ideal contoured (TIC) nozzles ". If the thrust-optimization is accentuated, the rapid sign inversion of the profile curvature just downstream of the throat induces a faster focusing of Mach lines, which generates an internal shock similar to the classical barrel shock forming in underexpanded propulsive jets. This focusing shock divides the core flow in two parts : a high speed and high momentum flow region around the nozzle centerline, and a high pressure level region near the wall. These highly thrust-optimized nozzles, as the Vulcain-type nozzles or the SSME nozzles [1], are called here "thrust-optimized contoured (TOC) nozzles ". Recent investigations [2-7] performed in the framework of the European "Flow Separation Control Device (FSCD)" program initiated by Cnes (the French Space Agency), tend to show that side loads 1 American Institute of Aeronautics and Astronautics (c)2001 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization. measured in overexpanded nozzles are strongly depending on the thrust-optimization level of the nozzle contour. As the side load phenomenon is caused by flow separation, side load theoretical models could be improved if we had a better understanding of the aerodynamics of overexpanded nozzles. This need has motivated the present experimental study. Test set-up, models and pressure instrumentation The tests have been performed in the Onera R2Ch blowdown wind tunnel [8], see Fig. 1, of the Fundamental and Experimental Aerodynamics Fig. 1 Nozzle model in the Onera R2Ch blowdown wind tunnel Two nozzle models, with different contours, have been tested. The first one has a classical bell profile, which results in a Truncated Ideal Contoured (TIC) nozzle. Figure 2 shows a photograph and a sketch of the TIC nozzle model. This one-body nozzle model, has a classical bell internal contour, a throat diameter of 20mm and an exit diameter of 90.9mm. The throatto-exit length is 148mm.